We observed that generally the embedding representation is very rich and information dense. For example, reducing the dimensionality of the inputs using SVD or PCA, even by 10%, generally results in worse downstream performance on specific tasks.
B. Sarwar, G. Karypis, J. Konstan, und J. Riedl. In Proceedings of The 6th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Workshop on Web Mining for E--Commerce -- Challenges and Opportunities (WEBKDD'00), (August 2000)
S. Cohen, K. Stratos, M. Collins, D. Foster, und L. Ungar. Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), Seite 223--231. Jeju Island, Korea, Association for Computational Linguistics, (Juli 2012)